The dataset collects data from an Android smartphone positioned in the chest pocket. Accelerometer Data are collected from 22 participants walking in the wild over a predefined path. The dataset is intended for Activity Recognition research purposes. It provides challenges for identification and authentication of people using motion patterns. Sampling frequency of the accelerometer: DELAY_FASTEST with network connections disabled.

Samples total







int, 1-22


classification, clustering


In this article, a novel technique for user’s authentication and verification using gait as a biometric unobtrusive pattern is proposed. The method is based on a two stages pipeline. First, a general activity recognition classifier is personalized for an specific user using a small sample of her/his walking pattern. As a result, the system is much more selective with respect to the new walking pattern. A second stage verifies whether the user is an authorized one or not. This stage is defined as a one-class classification problem. In order to solve this problem, a four-layer architecture is built around the geometric concept of convex hull. This architecture allows to improve robustness to outliers, modeling non-convex shapes, and to take into account temporal coherence information. Two different scenarios are proposed as validation with two different wearable systems. First, a custom high-performance wearable system is built and used in a free environment. A second dataset is acquired from an Android-based commercial device in a ‘wild’ scenario with rough terrains, adversarial conditions, crowded places and obstacles. Results on both systems and datasets are very promising, reducing the verification error rates by an order of magnitude with respect to the state-of-the-art technologies.


Downloaded from the UCI Machine Learning Repository on August 23, 2018.

Casale, Pierluigi, Oriol Pujol, and Petia Radeva. “Personalization and user verification in wearable systems using biometric walking patterns.” Personal and Ubiquitous Computing 16.5 (2012): 563-580.


yellowbrick.datasets.loaders.load_walking(data_home=None, return_dataset=False)[source]

Loads the walking activity dataset that is weill suited to clustering and multi-label classification tasks. The dataset contains multi-variate time series data with 149,332 real valued measurements across 22 unique walkers.

The Yellowbrick datasets are hosted online and when requested, the dataset is downloaded to your local computer for use. Note that if the dataset hasn’t been downloaded before, an Internet connection is required. However, if the data is cached locally, no data will be downloaded. Yellowbrick checks the known signature of the dataset with the data downloaded to ensure the download completes successfully.

Datasets are stored alongside the code, but the location can be specified with the data_home parameter or the $YELLOWBRICK_DATA envvar.

data_homestr, optional

The path on disk where data is stored. If not passed in, it is looked up from $YELLOWBRICK_DATA or the default returned by get_data_home.

return_datasetbool, default=False

Return the raw dataset object instead of X and y numpy arrays to get access to alternative targets, extra features, content and meta.

Xarray-like with shape (n_instances, n_features) if return_dataset=False

A pandas DataFrame or numpy array describing the instance features.

yarray-like with shape (n_instances,) if return_dataset=False

A pandas Series or numpy array describing the target vector.

datasetDataset instance if return_dataset=True

The Yellowbrick Dataset object provides an interface to accessing the data in a variety of formats as well as associated metadata and content.